2.2. Adrenomedullin

2.2.1. Discovery and gene cloning

AM was discovered and isolated from human pheochromocytoma extracts (Kitamura et al. 1993a). This 52-amino-acid peptide was observed to stimulate adenylyl cyclase activity in a platelet bioassay (Kitamura et al. 1993a). AM showed 27 % homology with the calcitonin gene-related peptide (CGRP) family (Fig. 3) and was found to have potent and long-lasting vasorelaxing effect (Kitamura et al. 1993a). The AM gene was cloned from a cDNA library from the human pheochromocytoma tissue (Kitamura et al. 1993b). As shown in Fig. 2., the human preproAM is 185 amino acids in length and is then processed to a 164-amino acid peptide, which is called proAM. The biologically active peptide is 52 amino acids in length and formed from proAM (Kitamura et al. 1993b). Suprisingly, it was found that the preproadrenomedullin contained a unique 20-residue sequence followed by Gly-Lys-Arg in the N-terminal region. This peptide was termed proadrenomedullin N-terminal 20 peptide (PAMP) (Kitamura et al. 1993b) and observed to have a dose-dependent hypotensive effect in rats (Kitamura et al. 1994b).

AM has a unique 6-residue ring structure and C-terminal amidation, which is similar to CGRP and amylin. The genomic DNA of human AM consists of 4 exons and 3 introns, and the 5´flanking region contains TATA, CAAT and GC boxes (Ishimitsu et al. 1994). The fourth exon codes for the mature form of AM (Ishimitsu et al. 1994). There are also multiple binding sites for activator protein-2 (AP-2) and cAMP regulated enhancer element (Ishimitsu et al. 1994). Southern blot analyses have indicated that the human AM gene is situated in a single locus on chromosome 11(Ishimitsu et al. 1994). Rat AM was cloned and it differs from human AM in only six positions and is 50 amino acids in length (Sakata et al. 1993). Human and rat PAMP are 20 amino acids in length and differ in three positions (Sakata et al. 1993).

2.2.2. AM receptors

Initially, the pharmacological actions of AM were thought to be mediated through the CGRP receptor, because of the significant structural homology between AM and CGRP and because some of the actions of AM were blocked by CGRP antagonist. However, today there are controversial results concerning AM receptor(s). Originally, AM was shown to generate increased production of cAMP from rat vascular smooth muscle cells (VSMC) (Ishizaka et al. 1994) and in the same study, specific binding sites for ¹²⁵I-AM were competitively inhibited by human AM, but not CGRP. However, in this experiment CGRP_8-37, a CGRP antagonist, competitively inhibited cAMP production by AM. The vasorelaxant effect of AM in cerebral arterioles and mesenteric arteries was also blocked by CGRP 8-37 (Mori et al. 1997, Lang et al. 1997, Berthiaume et al. 1995, Nuki et al. 1993). On the other hand, the blood pressure-lowering effect of infused AM was not blocked by CGRP_8-37 (Nandha et al. 1996, Haynes et al. 1995, Hjelmqvist et al. 1997). Moreover, the effect of AM to increase renal blood flow in isolated or in situ kidney was not blocked by CGRP_8-37 (Haynes et al. 1995, Gardiner et al. 1995). Owji et al. confirmed the presence of an abundant and specific binding site for AM in rat heart, lung, spleen, liver, skeletal muscle and spinal cord, with lesser binding in thyroid, adrenal, stomach, kidney and central nervous system (Owji et al. 1995). The binding was highest in rat heart and lung. In lung this binding was inhibited only by AM, whereas the binding in the heart was competitively inhibited by amylin and CGRP, suggesting regional variation in AM receptor specificity. In structure and activity studies (Eguchi et al. 1994) performed in rat VSMC cultures, human AM induced cAMP production that was augmented by GTP. Human AMs 1-52, 13-52 and 16-52 all inhibited equally binding of radiolabeled AM and caused similar stimulation of cAMP. C-terminal deletions of human AM exhibit reduced receptor binding and reduced cAMP production. Cysteine-to-cysteine six amino acid ring structure deletions of human AM exhibit no binding. These results confirm that AM receptors are coupled to adenylate cyclase via G-protein and that both the cyclic ring structure and C-terminal amidation are essential for full binding and activity.

The cDNA for AM receptor has been cloned from rat lung tissue (Kapas et al. 1995). This cDNA codes a polypeptide of 395 residues which possesses seven putative, α-helical transmembrane domains. Thus the AM receptor resembles other members of the G-protein linked receptor super-family. However, the degree of interaction by AM and CGRP with one others receptor has been unclear. In rat kidney mRNAs for AM and AM-receptor were found in renal vessels, glomeruli, and inner medullary collecting ducts. Ten days of feeding a low-salt (0.02%) or a high-salt diet (4%) did not change AM mRNA or AM-receptor mRNA in any kidney zone (Jensen et al. 1998). Most recently, McLatchie et al. (McLatchie et al. 1998) reported that receptor-activity-modifying proteins (RAMPs) are regulating the specificity of CGRP receptor. The CGRP receptor can function as either AM or CGRP receptor depending on which members of a new family of single-transmembrane-domain proteins, RAMPs, are expressed. RAMP1 presents the receptor at the cell surface as a mature glycoprotein and a CGRP receptor. RAMP2-transported receptors are core-glycosylated and function as AM receptors. Therefore, AM receptors and tissue distribution need to be studied further.

2.2.3. Biological effects of AM

Two hallmark actions of AM, hypotension and diuresis/natriuresis, suggest that adrenomedullin plays an important role in cardiorenal regulation (for reviews see Richards et al. 1996, Samson 1999). AM has important effects on renal function (for review see Samson 1999). In the brain AM inhibits water intake (Murphy & Samson 1995) and, in a physiologically relevant manner, salt appetite (Samson & Murphy 1997). Recently, AM antisense oligonucleotide treatment significantly lowered AM peptide content in the hypothalamic paraventricular nucleus and exaggerated the consumption of sodium, supporting a physiological role for AM in the central regulation of sodium homeostasis (Samson et al. 1999). Intrarenal infusion of AM in anesthetized dogs caused significant increases in renal blood flow (RBF), urine flow and sodium excretion in a dose-dependent manner without changes in heart rate or mean arterial pressure (MAP) (Ebara et al. 1994). These results indicated direct vascular and tubular effects of AM. In further studies, it was found that the effects of AM on RBF were mediated by nitric oxide (Miura et al. 1995). In anesthetized rats, intrarenal AM infusion had similar effects as in dogs, increases in RBF, arterial conductance, glomerular filtration rate (GFR), urine flow and sodium excretion (Haynes et al. 1995, Vari et al. 1996, Elhawary et al. 1995), and these effects were not blocked by a CGRP antagonist (Elhawary et al. 1995). Increases in RBF and urine flow were maintained also during systemic infusion of AM, which was associated with a marked decrease in MAP (Vari et al. 1996) and an increase in NO release in the kidney (Hirata et al. 1995).

Binding sites for AM and PAMP have been found in outer cortex and medulla of human adrenal gland (Nussdorfer et al. 1997). AM has been reported to inhibit angiotensin II- and potassium-stimulated, but not basal- or ACTH-stimulated, aldosterone release from cultured rat adrenal cells (Yamaguchi et al. 1996). Moreover, intravenous infusion of AM has been described to lower circulating cortisol levels and ACTH in sheep (Parkes et al. 1995).

Intravenous infusion of AM has been reported to cause prolonged hypotension in cat, rat, rabbit, sheep and humans (Ishiyama et al. 1995, Parkes et al. 1997, Hjelmqvist et al. 1997, Lainchbury et al. 1997, Parkes et al. 1995, Feng et al. 1994, Fukuhara et al. 1995, Nakamura et al. 1997, Parkes 1995, Shirai et al. 1997, Champion et al. 1997), largely through the generation of nitric oxide (NO) in the vasculature (Miura et al. 1995, Hirata et al. 1995, Feng et al. 1994). Bolus injections of AM (0.1-3.0 nmol/kg) in anesthetized SHR and Wistar-Kyoto (WKY) rats caused significant decreases in blood pressure (24-92 mmHg in SHRs vs 12-62 mmHg in WKY rats). In the same study, constant infusion of AM (0.03 nmol/min/kg) for 30 min reduced blood pressure 23 % and 20 % in SHR and WKY rats, respectively, without any effect on heart rate. In conscious rabbits, bolus injections of AM (10 and 3000 pmol/kg) dose-depently decreased MAP up to 27 % and the decrease in MAP was associated with an increase in heart rate (Fukuhara et al. 1995). In conscious SHR and SD rats, intravenous high dose (1670 and 5000 ng/kg) bolus injections of AM decreased blood pressure in both groups (He et al. 1995), while heart rate and cardiac output (CO) were increased. In the same study, using radioactive microspheres, AM infusion was shown to increase blood flow in the lungs, spleen, kidneys, adrenal glands and small intestine. The flow rates in brain and skin remained unchanged, whereas the flow rates were decreased in skeletal muscle and testes. In conscious sheep, 90 min AM infusion at 100 ng/min/kg reduced MAP by 12 mmHg, increased heart rate by 20 beats/min and CO 3 l/min. Moreover, plasma renin activity was elevated during AM infusion, whereas plasma aldosterone was not affected, and plasma norepinephrine levels fell (Charles et al. 1997). Recently, AM infusion at low (0.01 g/kg/min) and high (0.05 g/kg/min) doses was observed to exert diuresis and natriuresis without hypotension in experimental heart failure rats (Nagaya et al. 1999a). Moreover, in this study CO, increased suggesting a beneficial role for AM in heart failure. The increased heart rate is not secondary to the hypotension, because increased contractility was also observed in conscious sheep cardiac when pressure was maintained constant (Parkes 1995). In humans, intravenous infusion of AM (2 ng and 8 ng/min/kg) reduced blood pressure, but did not have any effect on urine volume and electrolyte excretion (Lainchbury et al. 1997). Intravenous injection of DNA constructs containing the human AM cDNA fused to the cytomegalovirus promoter, induced a long-lasting reduction in blood pressure in SHR and the maximal decrease in blood pressure was 22 mmHg (Chao et al. 1997).

The effect of AM has been studied in isolated organs (for reviews see Richards et al. 1996, Samson 1999). AM was reported to have a vasodilatory effect on perfused rat mesenteric vascular bed and this vasodilatation was inhibited by CGRP_8-37 (Nuki et al. 1993). In this study atropine and propranolol did not have any effect on AM-induced vasodilatation, suggesting that AM induced noncholinergic and nonadrenergic vasodilatation. AM has been shown to induce a dose-dependent vasorelaxation in basilar (Baskaya et al. 1995), mesenteric, coronary, renal and femoral arteries isolated from the dog (Nakamura et al. 1995). It is noteworthy that vasorelaxing effects were slightly greater in endothelium-intact arteries than in denuded arteries (Nakamura et al. 1995). In the rat isolated perfused kidney, AM caused a dose-dependent vasodilator response which was blocked dy CGRP_8-37 (Haynes et al. 1995, Gardiner et al. 1995). In isolated rat hearts bolus infusion of AM caused a dose-dependent and long-lasting coronary artery vasodilation, which was markedly attenuated by CGRP (8-37) (Entzeroth et al. 1995). Moreover, AM induced dose-depently an increase in developed tension in isolated perfused heart preparation, indicating a positive inotropic effect for AM (Szokodi et al. 1998). This effect was not altered by CGRP_8-37 or cAMP-dependent protein kinase inhibitor.

Both proliferative and antiproliferative effects of AM have been observed in vitro. In rat VSMCs, enhancement of thymidine incorporation and cell number by fetal calf serum is inhibited by AM and the inhibitory effect of AM was blocked by CGRP_8-37 (Kano et al. 1996). Furthermore, in cultured human coronary artery VSMCs, angiotensin-II stimulated cell migration was inhibited by AM (Kohno et al. 1997). These results suggest a role for AM in preventing pathological vascular remodeling. Moreover, AM blocked mitogen-stimulated mesangial cell proliferation in cell culture via the cAMP pathway (Chini et al. 1995). In contrast, in Swiss 3T3 fibroblast cell culture AM was observed to stimulate DNA synthesis and cell proliferation (Withers et al. 1996).

2.2.4. Gene expression of AM

AM mRNA has been found in several human tissues including adrenal medulla, cardiac ventricle, lung and kidney (Kitamura et al. 1993b). Using in situ hybridization, the expression of AM mRNA was examined in various tissues in rat and mouse. Intense expression of AM mRNA was observed in endometrium and epithelial cells lining the uterus and mouse adrenal medulla. Moderate levels of expression were detected in kidney glomerulus and cortical distal tubules, ovarian corpus luteum and follicles, epithelial cells lining the bronchioles, cardiac atrium and ventricle, posterior pituitary, stomach, small intestine, spleen and pancreas. Lower levels were observed in pulmonary alveoli, anterior pituitary and submandibular gland (Cameron & Fleming 1998). Rat and porcine vascular endothelial cells have been reported to synthetize and secrete AM (Sugo et al. 1994b). The most abundant transcription of AM is detected in rat endothelial cells, at an intensity 20 to 40 fold higher than that in the adrenal gland (Sugo et al. 1994a).

The changes in AM gene expression have been studied under different physiological and pharmacological stimuli in vitro and in vivo (for reviews see Richards et al. 1996, Samson 1999). By far the best-characterized activators of AM gene expression are cytokines and growth factors. In cultured VSMCs lipopolysaccharide (LPS), interleukin-1 (IL-1) and tumor necrosis factor alfa (TNF-α) stimulated AM gene expression. IL-1 and TNF-α were the most powerful, resulting in a 5 to 6-fold increase in AM gene expression at 14 hours (Sugo et al. 1995). This is most likely mediated via induction of the nuclear factor for interleukin-6 expression (NF-IL6) (Ishimitsu et al. 1998). Moreover, AM gene expression has been reported to be stimulated by IL-1 and TNF-α at 24 and 48 hours in both cultured rat cardiac myocytes and non-myocytes (Horio et al. 1998). In rats, LPS administration for 3 hours elevates AM gene expression 2 to 7-fold in aorta, lung, adrenal gland, skeletal and cardiac muscle, ileum/jejunum, brain and kidney (Shoji et al. 1995).

Adrenocortical steroids were observed to increase AM mRNA levels in cultured VSMCs (Imai et al. 1995, Minamino et al. 1995), and the increase in AM mRNA levels was enhanced in the presence of cycloheximide, whereas actinomycin D completely suppressed the AM gene expression (Minamino et al. 1995). These data indicated a rapid turnover of AM mRNA. Recently, Ando et al. showed that oxidative stress stimulates AM gene expression in cultured rat VSMCs at 24 hours (Ando et al. 1998). In contrast, shear-stress has been reported to down-regulate AM gene expression in cultured human aortic endothelial cells (Shinoki et al. 1998). In the rat model of cerebral ischemia, AM mRNA levels were reported to increase up to 20-fold (Wang et al. 1995). In hyperthyreoid rats, AM gene expression is increased in the lung compared to hypothyreoid rats (Murakami et al. 1998).

Cardiac AM gene expression responses to various stimuli have been studied in different animal models (for reviews see Richards et al. 1996, Samson 1999). Jougasaki et al. reported that in experimental congestive heart failure produced by rapid ventricular pacing in dogs, ventricular AM is activated during the progression of heart failure (Jougasaki et al. 1997). Furthermore, there was a positive correlation between left ventricular mass index and left ventricular AM concentrations, suggesting that ventricular AM expression may be increased by the ventricular hypertrophy in this model of heart failure (Jougasaki et al. 1997). Left ventricular AM mRNA levels were also significantly increased, 40 % by left coronary artery ligation in rats at 28 days (Kaiser et al. 1998), 28-34 % in hypertension of Dahl salt-sensitive rats on high-salt diet at 3 weeks (Shimokubo et al. 1996) and 25 % during aortic banding in SD rats at 1 day (Morimoto et al. 1999). The right ventricular AM mRNA and ir-AM levels were increased in rat experimental pulmonary hypertension (Shimokubo et al. 1995).

Although cardiac overload (Nishikimi et al. 1997a) and rapid ventricular pacing (Jougasaki et al. 1997) are known to alter the expression of the AM gene, it has not yet been established whether wall stretch acts directly or via local paracrine and autocrine factors liberated in response to hemodynamic load. Little is known about the AM gene expression in cardiac atria, although AM immunostaining was observed to be more intense in rat cardiac atria than ventricles (Jougasaki et al. 1995a). In addition, the time-course of cardiac AM transcription in the acute phase of cardiac overload is unclear.

2.2.5. Production and release of AM

The tissue distribution of ir-AM and AM mRNA have been widely examined in rat, man, pig and mouse (for reviews see Richards et al. 1996, Samson 1999, Eto et al. 1999). Ir-AM is detectable in multiple human tissues including adrenal medulla, heart, aorta, kidney, brain, lung, gastrointestinal organs, spleen, brain and thyroid (Ichiki et al. 1994, Satoh et al. 1997, Washimine et al. 1995a). The highest tissue concentrations of AM are found in the adrenal. Next in ranking is cardiac atrium, having only 3-4 % of the concentration present in adrenal (Ichiki et al. 1994). In the rat, ir-AM is present in various tissues, and in high concentrations in the adrenal gland, lung and cardiac atrium (Sakata et al. 1994). The AM immunostaining in the atria of canine heart is more intense than in the ventricles (Jougasaki et al. 1995a).

In humans, AM is circulating in the low pg/ml range (Kitamura et al. 1994a, Sato et al. 1995). The plasma half-life of AM is estimated to be about 22 minutes in humans (Meeran et al. 1997). In humans, the concentration of circulating AM has been shown to be increased in patients with congestive heart failure (Jougasaki et al. 1995b, Kato et al. 1996, Kobayashi et al. 1996b), essential hypertension (Kitamura et al. 1994a, Ishimitsu et al. 1994, Kohno et al. 1996b), chronic renal failure (Ishimitsu et al. 1994, Washimine et al. 1995b), pulmonary hypertension (Shimokubo et al. 1995, Nishikimi et al. 1997b), myocardial infarction (Miyao et al. 1998, Yoshitomi et al. 1998), septic shock (Hirata et al. 1996, Nishio et al. 1997), thyreotoxicosis (Taniyama et al. 1996) and acute asthma (Kohno et al. 1996b). In hypertension plasma AM is increased and the increase has been reported to be pronounced in hypertension with target organ damage (Ishimitsu et al. 1994). Furthermore, it has been reported that hypertension-induced plasma AM levels did not fall with effective treatment of hypertension (Kohno et al. 1996b). In chronic renal impairment plasma AM levels are 2 to 3-fold compared to normal control values depending of the severity of renal impairment (Ishimitsu et al. 1994). In human heart failure plasma AM levels have been reported to be elevated (Jougasaki et al. 1995b, Kato et al. 1996, Kobayashi et al. 1996b) and the plasma AM levels decreased with effective treatment (Kita et al. 1998). Moreover, Jougasaki et al. demonstrated that immunohistochemical staining of AM is significantly increased in the failing human ventricular myocardium compared with the normal human ventricle (Jougasaki et al. 1995b). Failing human ventricular myocardium has been shown to secrete AM (Jougasaki et al. 1996). In acute myocardial infarction (AMI), plasma levels of AM rose to a peak after 48-72 h and remained above preinfarction level for up to 3 weeks (Kobayashi et al. 1996a). Moreover, plasma AM on day 2 after AMI was found to be associated with long-term mortality, and AM could thus be used as a prognostic indicator (Nagaya et al. 1999b). In rats, circulating AM is increased in hypertension of Dahl salt-sensitive rats on a high-salt diet (Shimokubo et al. 1996), in experimental pulmonary hypertension (Shimokubo et al. 1995) and in heart failure rats (Nishikimi et al. 1997a).